[Jmap] questions/comments on draft-ietf-jmap-webpush-vapid-01

Jim Fenton <fenton@bluepopcorn.net> Sat, 06 April 2024 20:55 UTC

From: Jim Fenton <fenton@bluepopcorn.net>
To: Daniel Gultsch <daniel@gultsch.de>, IETF JMAP Mailing List <jmap@ietf.org>
Date: Sat, 06 Apr 2024 13:55:39 -0700
Message-ID: <EAD39284-16CF-4E5B-B8DB-200475FC4104@bluepopcorn.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"; markup="markdown"
Content-Transfer-Encoding: quoted-printable
Archived-At: <https://mailarchive.ietf.org/arch/msg/jmap/sBjZ7zE7Nqq6xs8cyOXkPCrVQSE>
Subject: [Jmap] questions/comments on draft-ietf-jmap-webpush-vapid-01
Precedence: list

I’m unfortunately not familiar with webpush at all, but I read through this draft and one thing stuck out. In Section 3, the applicationServerKey was described as a P-256 public key, which made me ask myself, “what about algorithm agility?”  It also said that the public key was encoded in URL-safe base64 [RFC4648].

I went over to RFC 8292 (VAPID for Web Push), and there the public key parameter seems to be in the form of a JSON Web Key (JWK) [RFC 7517]. There the key type and curve (for elliptical) seems to be specified along with the public key. That would answer my concern about algorithm agility, and it probably also makes security sense to ensure the key is only used with the intended algorithm. Should the applicationServerKey instead be specified to be a (probably base64 encoded) JWK?

Also, is there any provision for key rollover, for example allowing either of two keys to be used during the transition between one signing key and another? Or is that not needed for some reason?

-Jim (as participant)

[Jmap] questions/comments on draft-ietf-jmap-webp… Jim Fenton